View Single Post
Old 12-07-2017, 08:36 AM   #1
Location: Sheffield, UK

Join Date: Feb 2012
Posts: 15
Default BiSeq - How to handle covariates?


I am attempting to use BiSeq (Hebestreit et al. 2013) to delineate DMRs among treatment groups in Bismark aligned RRBS data

I have made a successful preliminary run through of the BiSeq workflow using my data but now I would like account for the effects of a covariate on DMR identification. BiSeq can apparently be made to work with covariates but I can't find any advice in the user guide as to how to go about doing this

I have tried modifying the betaRegression() formula parameter to take into account metadata (where the covariate values are stored) as a covariate:

RSbetaResults <- betaRegression(formula = ~group + metadata,
link = "probit",
object = RSpredictedMeth,
type = "BR")

Unfortunately, this did not work

Any suggestions/advice would be gratefully received
pig_raffles is offline   Reply With Quote